Nociceptin inhibits calcium channel currents in a subpopulation of small nociceptive trigeminal ganglion neurons in mouse



  • 1The effects of nociceptin/orphanin FQ (N/OFQ) and opioid receptor agonists on voltage-activated calcium channel currents (ICa) were examined in acutely isolated mouse trigeminal ganglion neurons using whole-cell patch-clamp recordings. These effects were correlated with responses of the neurons to capsaicin and binding of Bandeiraea simplicifolia isolectin B4 (IB4).
  • 2Trigeminal neurons were divided into two populations based on the presence (type 2) or absence (type 1) of a prominent T-type ICa. N/OFQ potently (EC50 of 19 nm) inhibited high-voltage-activated (HVA) ICa in most (82 %) small (capacitance < 12 pF) type 1 neurons, but few (9 %) larger (> 12 pF) type 1 neurons. N/OFQ inhibited ICa in few (23 %) type 2 cells, and did not affect the T-type ICa in any cell.
  • 3The μ-opioid agonists DAMGO and morphine inhibited ICa in most type 1 neurons, more often (95 %versus 77 %) in the small cells. The inhibition of ICa by DAMGO and morphine was more efficacious in small versus large type 1 neurons. μ-Opioids did not inhibit ICa in type 2 neurons.
  • 4Most small type 1 neurons were sensitive to capsaicin (93 %) and bound IB4 (86 %). Fewer larger type 1 neurons responded to capsaicin (30 %) or bound IB4 (58 %). Type 2 neurons did not respond to capsaicin, although some bound IB4 (35 %).
  • 5Thus, N/OFQ preferentially inhibits HVA ICa in a subpopulation of small nociceptive trigeminal ganglion neurons that is also highly sensitive to μ-opioid agonists.

Nociceptin (N/OFQ; Meunier et al. 1995; also called orphanin FQ; Reinsheid et al. 1995) is an endogenous peptide that selectively acts on the opioid-like receptor (ORL1; Mollereau et al. 1994). Although initially postulated to function as a pro-nociceptive peptide (Meunier et al. 1995), the role of N/OFQ in pain modulation is complex. Either pro-nociceptive or anti-nociceptive effects can be evoked by N/OFQ at different sites in the central nervous system (for reviews, see Meunier, 1997; Xu et al. 2000).

The actions of N/OFQ on excitability of primary afferent neurons depend on location of application of the peptide and the type of nerve examined (Giuliani et al. 2000). N/OFQ presynaptically inhibits excitatory postsynaptic potentials and currents evoked by stimulation of the entry zone of primary afferent nerves in the rat spinal cord dorsal horn neurons (Lai et al. 1997; Liebel et al. 1997). Similarly, N/OFQ inhibits C-fibre-evoked responses in rat dorsal horn in vivo (Carpenter et al. 2000) and responses of nucleus trigeminalis caudalis neurons to noxious stimuli (Wang et al. 1999), suggesting an anti-nociceptive action. Although high concentrations of N/OFQ also depress non-noxious excitatory transmission in the rat dorsal horn (Faber et al. 1996) and nucleus trigminalis caudalis (Wang et al. 1999), it is not clear whether or not these effects are caused by an inhibitory action on primary afferent terminals, or directly on dorsal horn neurons (e.g. Jennings, 2001). Predominantly anti-nociceptive responses to intrathecal application of N/OFQ have been reported (Xu et al. 2000), consistent with presynaptic inhibition of C-fibre nociceptive neurotransmission. In contrast, direct application of N/OFQ to peripheral receptive fields of primary afferent nerves of rats and mice evokes nociceptive behavioural responses (Inoue et al. 1998) and excitation of dorsal horn neurons (Carpenter et al. 2000). In isolated tissues, N/OFQ generally (but not universally) inhibits sensory nerve-evoked activity (Giuliani et al. 2000).

It is not known whether ORL1 receptors are expressed by a subpopulation of nociceptive primary afferent neurons, non-nociceptive neurons, or both. The ORL1 receptor has been identified by in situ hybridization in large neurons in dorsal root ganglia (Neal et al. 1999). N/OFQ inhibits voltage-dependent calcium channel currents (ICa) in some rat dorsal root ganglion (DRG) neurons, a proportion of which also respond to morphine (Abdulla & Smith, 1997, 1998; Zhang et al. 1998). However, apart from noting that DRG cells of all sizes responded to N/OFQ, other characteristics of the neurons that respond to N/OFQ have not been identified. In the present study, we examined the effects of N/OFQ on low-voltage-activated (LVA) and high-voltage-activated (HVA) ICa in mouse trigeminal ganglion neurons. High sensitivity of HVA ICa to N/OFQ was restricted to a subpopulation of small sensory neurons that express VR1, isolectin B4 (IB4) binding and highly efficacious responses to μ-opioid agonists. These results suggest that the responsiveness of trigeminal ganglion neurons to N/OFQ is restricted to a subset of nociceptors.


C57BJ/16 hybrid mice (4–6 weeks old) of either sex were used for this study. All experiments were conducted according to protocols approved by the Animals Ethics Committee of the University of Sydney, Sydney, NSW, Australia. Mice were anaesthetized with halothane (4 %) and decapitated, and the trigeminal ganglia were removed and placed in ice-cold physiological saline containing (mm): NaCl, 126; KCl, 2.5; CaCl2, 2.5; MgCl2, 1.2; NaH2PO4, 1.2; NaHCO3, 24; and glucose, 10, gassed with 95 % O2-5 % CO2. Cells were prepared using a simplified version of the methods outlined in Eckert et al. (1997). Ganglia were cut up with iridectomy scissors and incubated at 32–34 °C for 30 min in physiological saline. The ganglia pieces were then transferred to oxygenated Hepes buffered saline (HBS) containing 20 units ml−1 papain and incubated at 32–34 °C for 15–20 min. The HBS contained (mm): NaCl, 154; KCl, 2.5; CaCl2, 2.5; MgCl2, 1.5; Hepes, 10; glucose, 10; pH 7.2 (NaOH), 330 ± 5 mosmol l−1. The digestion was terminated with addition of HBS containing 1 mg ml−1 bovine serum albumin (BSA) and 1 mg ml−1 trypsin inhibitor. Minced ganglia were washed free of enzyme and enzyme inhibitors with room-temperature HBS. Cells were released by gentle trituration through decreasing bore, silanized Pasteur pipettes with fire-polished tips. The cells were plated onto plastic culture dishes and kept at room temperature in HBS. Cells remained viable for up to 10 h after dissociation and could be cultured overnight.

Pertussis toxin treatment

Trigeminal ganglion cells were isolated as described above. Neurons in HBS were left to attach to plastic culture dishes before changing media to L15 Leibovitz (Sigma) containing 5 % fetal bovine serum (Gibco), penicillin-streptomycin 50 units-5 μg ml−1 and incubated at 25 °C. Pertussis toxin (PTX; 100 ng ml−1) was added to the culture media prior to overnight incubation.

IB4 treatment

Cells were pre-treated with 10 μg ml−1 fluorescein isothiocyanate (FITC)-conjugated Bandeiraea simplicifolia isolectin B4 (Sigma) for 20 min at 22–24 °C, as described by Stucky & Lewin (1999). Neurons were washed twice with HBS before fluorescence was examined on an inverted microscope. Recordings were then made as described below.

Electrophysiological recording

Ionic currents from mouse trigeminal neurons were recorded in the whole-cell configuration of the patch-clamp method (Hamill et al. 1981) at room temperature (22–24 °C). Dishes were continually perfused with HBS. For isolating ICa, the extracellular solution containing (mm): tetraethylammonium chloride (TEACl), 140; CsCl, 2.5; CaCl2, 2.5; Hepes, 10; MgCl2, 1; glucose, 10; pH 7.2 (CsOH) 330 ± 5 mosmol l−1 was used. In a few early experiments to determine the sensitivity of the trigeminal neurons to opioid receptor agonists, 5 mm BaCl2 was used as the charge carrier instead of 2.5 mm CaCl2 (see Fig. 5B and Fig. 6A). The intracellular pipette solution contained (in mm): CsCl, 120; Hepes, 10; EGTA, 10; CaCl2, 2; MgATP, 5; Na2GTP, 0.2; NaCl, 5; pH 7.3 (CsOH); 285 ± 5 mosmol l−1. For recording capsaicin-induced currents, the extracellular solution contained (mm): NaCl, 80; TEACl, 60; CsCl, 2.5; MgCl2, 5; Hepes, 10; glucose, 10; pH 7.2 (with CsOH), 330 ± 5 mosmol l−1. All reported membrane potentials have been corrected for a liquid junction potential of either −7 mV (ICa recordings) or −4 mV (capsaicin current recordings).

Figure 5.

A, a representative current trace from a type 1 neuron in the presence or absence of 1 μm capsaicin. Currents were elicited by the ramp protocol shown below the example traces. B, examples of ICa currents recorded before, during and after application of DAMGO in a type 1 neuron. C, N/OFQ inhibits HVA ICa (at −7 mV) preferentially in small (< 12 pF, open symbols) versus larger (> 12 pF, filled symbols) type 1 neurons. Each data point represents the inhibition of HVA ICa by N/OFQ (300 nm to 1 μm) in a type 1 neuron. Circles represent neurons that responded and triangles represent those that did not respond to N/OFQ (see text for definition of a positive response). D, small type 1 neurons could be distinguished from large ones on the basis of responses to N/OFQ, DAMGO, capsaicin and binding of IB4. Inhibition of HVA ICa (at −7 mV) was determined in the presence or absence of 1–3 μm DAMGO or 300 nm to 1 μm N/OFQ. IB4 sensitivity was determined using fluorescence microscopy after incubation with FITC-conjugated IB4. Bars represent percentage of cells that were sensitive to the treatment. Large type 1 neurons had significantly lower frequencies of response to N/OFQ, DAMGO, capsaicin and IB4 than small type 1 neurons. *P < 0.05, **P < 0.0012-test, small versus large type 1 neurons).

Figure 6.

Small type 1 neurons are more sensitive to μ-opioids than large type 1 neurons

A, an example trace of HVA ICa inhibited by morphine in a type 1 neuron. Inhibition was blocked in the presence of the μ-opioid antagonist CTAP. B, DAMGO inhibited HVA ICa more potently and with greater maximal effect in small versus large type 1 neurons (at least 6 cells per point; ANOVA, P < 0.05). C, morphine inhibited HVA ICa with greater maximal effect in small versus large type 1 neurons (n= 5–8 cells per point, ANOVA, P < 0.02).

Recordings were made using an EPC-9 patch-clamp amplifier and corresponding PULSE software from HEKA Electronik (Lambrecht/Pfalz, Germany) or an Axopatch ID amplifier (Axon Instruments, Foster City, CA, USA) using pCLAMP acquisition software (v. 5.5, Axon Instruments). Currents were sampled at 20–50 kHz and recorded on hard disk for later analysis. Patch pipettes were pulled from borosilicate glass (AM Systems, Everett, WA, USA). The pipette input resistance ranged between 0.7 and 1.5 MΩ. The capacitance of individual cells ranged between 4 and 50 pF with a series resistance between 1 and 5 MΩ. Series resistance compensation of at least 80 % was used in all experiments. Capacitance transients were compensated automatically using a built-in procedure of the HEKA amplifier or by using the manual compensation on the Axopatch-1D. Leak current was subtracted online using a P/8 protocol unless otherwise noted (see Fig. 4). The mean current settling time (from 12 randomly selected traces) was 0.6 ± 0.05 ms.

Figure 4.

Inhibition of HVA ICa by N/OFQ is mediated by βγ subunits released from Gi/o proteins

A, pre-pulse relief of inhibition indicates βγ-subunit-mediated inhibition. Neurons were initially stepped from −87 to −7 mV (S1). Neurons recovered for 100 ms (approximately 80 ms has been removed for clarity, as marked by bars on the trace and voltage protocol diagram) before a conditioning pre-pulse to +123 mV for 70 ms was applied. A second test pulse (S2) was elicited by stepping the membrane potential from −87 to −7 mV. Inhibition of HVA ICa by 300 nm N/OFQ was greatly reduced in S2 as compared to S1. Leak subtraction was not used for this experiment. B, trigeminal neurons were incubated overnight at 25 °C with or without 100 ng ml−1 PTX. The PTX treatment abolished the inhibition of HVA ICa by 300 nm N/OFQ or 1 μm DAMGO. Bars represent the mean ±s.e.m. HVA ICa inhibition occurred in at least six cells per condition.

Peak HVA ICa in each cell was determined by stepping the membrane potential from a holding potential of −87 mV to between −67 and +63 mV, for 30 ms, in 10 mV increments, unless otherwise indicated. Following this procedure, the test current was evoked (−87 to −3 mV) every 30 s and monitored for current stability before drugs were applied. Cells were rejected if the current increased or decreased by more than 2 % in the first 90 s after running the I–V protocol. Cells were exposed to drugs via a series of flow pipes positioned above the cells. The inhibition by drugs was quantified by measuring the current isochronically from the peak of the control current in the presence and absence of the drug.

Data analysis

All data are expressed as the mean ±s.e.m. unless otherwise indicated. Concentration-response data were pooled for each group and fitted to a logistic equation using the software package GraphPad Prism v. 3. Where noted, significant differences between means were tested, using paired or unpaired two-tailed Student's t tests. Differences between frequency data were tested using χ2 tests. When comparing differences in the concentration-response curves of DAMGO or morphine in small type 1 or large type 1 neurons, a one-way analysis of variance (ANOVA) was used post hoc in conjunction with Bonferroni's corrected t test.

Drugs and chemicals

DAMGO ([Tyr-d-Ala-Gly-MePhe-Gly-ol]enkephalin), N/OFQ (Phe-Gly-Gly-Phe-Thr-Gly-Ala-Arg-Lys-Ser-Ala-Arg-Lys-Leu-Ala-Asn-Gln), CTAP (d-Phe-Cys-Tyr-d-Trp-Arg-Thr-Pen-Thr-NH2) and ω-conotoxin GVIA were from Auspep (Melbourne, Australia). Capsaicin was from Tocris Cookson (Bristol, UK). Morphine HCl was from Glaxo (Hertfordshire, UK). Nimodipine, naloxone hydrochloride and U69593 were from Research Biochemicals International (Natick, MA, USA). Buffer salts were from BDH Australia or Sigma Australia. ω-Agatoxin VI was from Peptide Institute Inc. (Osaka, Japan). Papain was from Worthington Biochemical Corporation (Freehold, NJ, USA). BSA, trypsin inhibitor (chicken egg white ovomucoid, Type II-O), pertussis toxin, IB4, DPDPE ([d-Pen2,5]-enkephalin) and CdCl2 were from Sigma Australia. The compound J-113397 was graciously donated by Banyu Pharmaceuticals (Tokyo, Japan).


Mouse trigeminal neurons can be divided into two subpopulations based on electrophysiological and pharmacological criteria

Mouse trigeminal neurons were readily subdivided into two groups based on several pharmacological and electrophysiological criteria. Initial studies focused on the sensitivity of the neurons to opioid agonists, and to the vanilloid receptor 1 (VR1) agonist capsaicin. When recording ICa, we routinely obtained complete current- voltage relationships to evaluate the integrity of the voltage clamp. It became evident that there were two populations of trigeminal neurons that could be distinguished by the presence (type 2) or absence (type 1) of a prominent low-voltage-activated ‘T-type’ calcium current (LVA ICa; Fig. 1). The absence of LVA ICa was also highly correlated with sensitivity to the VR1 agonist capsaicin, and with inhibition of HVA ICa by opioid agonists. These properties are discussed in more detail below, but the strong correlation among several phenotypic markers enabled us to classify cells based on ICa current-voltage relationships.

Figure 1.

Calcium current characteristics of two populations of mouse trigeminal neurons

A and B, illustration of ICa elicited by steps from −87 to −47 and −7 mV for each cell type: type 1, neuron without T-type ICa (A); and type 2, neuron with T-type ICa (B). C, a current-voltage relationship of a neuron without LVA ICa (type 1 cell). D, a current-voltage relationship of a neuron with LVA ICa (type 2 cell). E, a frequency histogram of cell size in the two populations of neurons. Cells with LVA ICa (filled bars) are small, whereas cells without LVA ICa (open bars) have a greater range of cell sizes. F, Ni2+ inhibits LVA ICa more potently than Cd2+.

Cells were classified as type 2 sensory neurons if the current amplitude at −47 mV was at least 10 % of the peak current, which was usually between −7 and +3 mV, although the amplitude of the LVA ICa at −47 or −37 mV was greater than the peak HVA ICa in a few cells. Cells classified as type 1 sensory neurons exhibited ICa at −47 mV that was less than 10 % of the peak current (Fig. 1A and C). The peak ICa density of type 2 neurons was low compared to type 1 neurons. The peak current density was 151 ± 7 pA pF−1 (n= 148) for type 1 neurons and 70 ± 4 pA pF−1 (n= 63) for type 2 cells.

The nature of the LVA channels in type 2 neurons was further examined by evoking the currents from holding potentials of −100 mV to reduce steady-state inactivation of LVA T-type channels. ICa was first activated by steps to −70 mV (two out of 29) or −60 mV (27 out of 29) in all type 2 cells examined from a holding potential of −100 mV. In these cells, ICa at the test potential of −60 mV represented 13 ± 1 % of the peak inward ICa. The amplitude of ICa evoked by a step to −40 mV was 64 ± 3 % of the peak inward ICa in these cells. The sensitivity of the LVA ICa to Ni2+ and Cd2+ was determined by repetitively stepping type 2 neurons from −100 to −40 mV in the presence or absence of increasing concentrations of the cations. Ni2+ inhibited ICa at −40 mV with an EC50 of 48 μm (pEC50= 4.4± 0.1, Hill coefficient, nH= 0.8 ± 0.2; Fig. 1F), while Cd2+ inhibited the LVA ICa with an EC50 of 130 μm (pEC50= 3.9 ± 0.02, nH= 1.2 ± 0.05; Fig. 1F). The relatively negative reversal potential of the ICa in type 2 neurons, apparent in Fig. 1B, most probably reflects the predominance of LVA channels in these cells since the channels thought to comprise LVA ICa (α1G, α1H and α1I, now known as CaV3.1, CaV3.2 and CaV3.3, respectively; Ertel et al. 2000) have reversal potentials at relatively negative potentials when compared with HVA ICa (e.g. Serrano et al. 1999; McRory et al. 2001).

N/OFQ inhibits HVA ICa in a subpopulation of type 1 mouse trigeminal neurons

N/OFQ inhibited HVA ICa in 53 % (52 out of 98) of type 1 neurons (Fig. 2 and Fig. 5C and D). Neurons were defined as N/OFQ sensitive if they showed a reversible inhibition of HVA ICa of greater then 10 % when exposed to maximally effective concentrations of N/OFQ (300 nm to 1 μm). In contrast to type 1 cells, high concentrations of N/OFQ (300 nm to 1 μm) produced only a small inhibition of HVA ICa (4 ± 1 %, Fig. 2C and D) in type 2 neurons. In the 23 % (8/35) of type 2 neurons responsive to N/OFQ (300 nm), the inhibition was 15 ± 5 %. This concentration of N/OFQ also failed to inhibit LVA ICa evoked by depolarization from −87 to −47 mV in any of the five type 2 cells tested in which N/OFQ (Fig. 2C and D), produced more than 10 % inhibition of HVA ICa (Fig. 2C and D). In type 2 neurons where N/OFQ (300 nm to 1 μm) did not inhibit HVA ICa, N/OFQ also had no effect on LVA ICa (n= 16).

Figure 2.

N/OFQ reversibly inhibits HVA ICa in sensory neurons

A, 300 nm N/OFQ reversibly inhibits HVA ICa in sensory neurons without LVA ICa (type 1). HVA ICa was elicited by stepping the membrane potential from −87 to −7 mV. B, a representative example of a current-voltage relationship of a type 1 neuron in the presence (triangles) and absence (squares) of N/OFQ (300 nm). C, N/OFQ produced a small inhibition of HVA ICa, but not LVA ICa in some neurons with ‘T-type’ICa (type 2). LVA ICa was activated by stepping membrane potential from −87 to −47 mV, HVA ICa was activated by a step to −7 mV. Approximately 60 ms of the trace, as marked by bars, has been removed for clarity. D, a current-voltage relationship of a representative type 2 neuron in the presence (triangles) or absence (squares) of 300 nm N/OFQ.

To determine current-voltage (I–V) relationships for HVA ICa, membrane potential was stepped from −87 mV to potentials between −67 and + 53 mV. The inward current in most cells was detectable at approximately −30 mV, and reached a peak between −7 and +3 mV (Fig. 1B). Inhibition of HVA ICa in type 1 neurons by 300 nm N/OFQ occurred over a range of membrane potentials (Fig. 2B, see also below).

Inhibition of HVA ICa by N/OFQ in type 1 neurons was concentration dependent with an EC50 of 19 nm (pEC50= 7.7± 0.15) and maximum inhibition of 37 ± 6 % (Fig. 3A). Inhibition of HVA ICa by 100 nm N/OFQ was abolished by the non-peptidergic, selective ORL1 antagonist, J113397 (300 nm; Fig. 3B; Ozaki et al. 2000). Prior to co-application of N/OFQ and J113397, all neurons were determined to be responsive to N/OFQ (n= 11). J113397 applied alone had no effect on HVA ICa. Inhibition of HVA ICa by 300 nm N/OFQ was not reduced by the opioid receptor antagonist naloxone (1 μm; n= 6; data not shown).

Figure 3.

N/OFQ inhibits HVA ICa via activation of ORL1 receptors

A, inhibition of HVA ICa by N/OFQ is concentration dependent, with an EC50 of 19 nm. Each data point represents mean HVA ICa inhibition of at least six type 1 neurons. Error bars represent the s.e.m.B, inhibition of HVA ICa by 100 nm N/OFQ was abolished in the presence of 300 nm J113397. J113397 (300 nm) was applied to cells for 1 min before co-application of J113397 (300 nm) and N/OFQ (100 nm).

Mechanism of N/OFQ inhibition of HVA ICa in small type 1 neurons

The role of G-protein βγ subunits in the inhibition of calcium channels by N/OFQ was investigated using a pre-pulse protocol (Fig. 4A). In these experiments, HVA ICa was evoked by two test steps to 0 mV, separated by 190 ms. The second test step (S2) was preceded by a conditioning step to +123 mV for 70 ms (see voltage protocol diagram in Fig. 4A). In the presence of 300 nm N/OFQ, the amplitude evoked by the first test step (S1) was less than the amplitude evoked by the second test step (S2) that followed the conditioning step to +123 mV (n= 6; Fig. 4A, Table 1). Inhibition of HVA ICa was significantly reduced from 30 ± 3 % in S1 to 15 ± 4 % in S2 (P < 0.001). HVA ICa activation was examined by calculating the 0 to 50 % rise time for control and N/OFQ-treated cells (Table 1). In the presence of 300 nm N/OFQ, the 0 to 50 % rise times were significantly slowed when compared to the rise times of the control in S1 (P < 0.05). This slowing of activation of HVA ICa observed in the presence of N/OFQ is consistent with G-protein βγ-mediated inhibition of HVA ICa.

Table 1. Effects of a depolarizing conditioning step on N/OFQ modulation of HVA ICa in small type 1 mouse trigeminal neurons
 ControlN/OFQ (300 nm)
  1. Trigeminal neurons were stepped twice from −87 to −3 mV with a 70 ms conditioning step to +123 mV preceding the second step (see Fig. 4). S1, first test step; S2, second test step. Values represent the mean and s.e.m. The S2:S1 ratio and 0–50 % rise times were significantly greater on indicated values (Student's paired, two-tailed t test, n= 6). *P < 0.05, **P < 0.01.

Ratio S2:S1 peak amplitude0.9 ± 0.041.2 ± 0.07 **
0–50 % rise time of S1 (ms)0.9 ± 0.061.6 ± 0.05 *
0–50 % rise time of S2 (ms)1.0 ± 0.051.1 ± 0.06

To examine the nature of the G-protein involved in N/OFQ inhibition of HVA ICa, neurons were treated overnight with 100 ng ml−1 pertussis toxin (PTX), a selective inactivator of Gi/o coupling. Inhibition of HVA ICa by 300 nm N/OFQ was abolished in PTX-treated neurons (Fig. 4B). Inhibition of HVA ICa by 300 nm N/OFQ in the control neurons was unaffected by the overnight incubation (Fig. 4B). The inhibition of ICa by 1 μm DAMGO was also abolished by overnight treatment with PTX.

Type 1 sensory neurons can be divided into two subpopulations based on sensitivity to N/OFQ and cell size

In type 1 cells, N/OFQ selectively inhibited HVA ICa in small-sized neurons (Fig. 5C), particularly those with a capacitance of less than 12 pF; 82 % (49 out of 60) of type 1 neurons smaller than 12 pF were sensitive to N/OFQ versus 9 % (four out of 45) of larger type 1 neurons (χ2= 56.2, P < 0.001). A cell capacitance of 12 pF corresponds to a diameter of 20 μm (assuming that the cell surface is a smooth sphere and specific membrane capacitance of 1 μF cm−2). This estimate was confirmed by a comparison of cell diameter, measured using the microscope eyepiece graticule, with cell capacitance, determined using the manual capacitance compensation of the Axopatch-1D. Out of 25 cells in which both these measurements were made, 16 had an apparent diameter between 18 and 22 μm (19 ± 3 μm). The corresponding mean capacitance for these 16 cells was 11.4 ± 0.5 pF. To further classify the properties of type 1 neurons, we defined small neurons as those with a capacitance of less than 12 pF and large neurons as those with capacitance greater than or equal to 12 pF.

Small type 1 neurons express properties typical of nociceptors

Type 1 trigeminal neurons exhibited properties typical of nociceptors. Sensitivity to the vanilloid receptor agonist capsaicin was examined by holding the cells at −4 mV and ramping the membrane potential from +36 to −64 mV every 10 s. Application of capsaicin (1 μm) resulted in a large (1.2 ± 0.3 nA at +36 mV), reversible outward current in approximately 65 % of type 1 neurons (Fig. 5A). All but one of the type 2 neurons tested were capsaicin insensitive (see Fig. 5D).

μ-Opioid agonists inhibited ICa in type 1 but not type 2 neurons (Fig. 5B and D). DAMGO and morphine inhibited current through ICa via μ-opioid receptors. Inhibitions of ICa by 1 μm DAMGO (38 ± 6 %, n= 6) and 10 μm morphine (38 ± 5 %, n= 14) were completely blocked by the selective μ-opioid receptor antagonist CTAP (1 μm; n= 6, data not shown). Furthermore, the δ-opioid selective agonists DPDPE (1 μm; n= 13) and deltorphin II (1 μm; n= 9) as well as the κ-opioid selective agonist U69,593 (1 μm; n= 13) did not modulate ICa in any type 1 or type 2 neuron tested (data not shown).

Greater proportions of small than large type 1 neurons were sensitive to μ-opioids and capsaicin, and displayed IB4 binding (Fig. 5D). HVA ICa was inhibited by 1 or 3 μm DAMGO in 95 % (42/44) of small type 1 versus 77 % (24/31) of large type 1 neurons (Fig. 5D, χ2= 4.6, P < 0.05). In contrast, DAMGO (1–3 μm) did not inhibit HVA ICa of any type 2 neurons (0 out of 21, Fig. 5D). Application of capsaicin (1 μm) produced current in 93 % (42 out of 45) of small type 1 neurons, but only 38 % (14 out of 37) of large type 1 neurons (Fig. 5D, χ2= 29.0, P < 0.001). Only one out of 32 type 2 neurons tested responded to 1 μm capsaicin. IB4 has previously been used as a marker for labelling subpopulations of nociceptors in living neurons (Stucky & Lewin, 1999). FITC-IB4 labelled nearly all small type 1 neurons (19 out of 20, 95 %), but only 58 % of large type 1 neurons (7 out of 13, Fig. 5D, χ2= 7.7, P < 0.05). A proportion of (6 out of 17, 35 %) of type 2 neurons was labelled with FITC-IB4 (Fig. 5D).

DAMGO inhibited HVA ICa of small type 1 sensory neurons with an EC50 of 0.2 μm (pEC50 6.7 ± 0.1) to a maximum of 58 ± 7 % (Fig. 6B). DAMGO inhibited ICa less potently in large type 1 sensory neurons with an EC50 of 0.5 μm (pEC50= 6.2± 0.1) to a maximum of 46 ± 3 % (Fig. 6B). Morphine inhibited HVA ICa in small type 1 neurons EC50 0.6 μm (pEC50 6.2 ± 0.1) and large type 1 neurons with similar potency (EC50= 0.7μm, pEC50= 6.1± 0.1; Fig. 6C). However, maximum inhibition of HVA ICa caused by morphine was greater in small type 1 (57 ± 3 %) than large type 1 sensory neurons (31 ± 8 %; Fig. 6A and C, P < 0.02).

Characterization of HVA ICa types expressed in mouse trigeminal sensory neurons

Differential responsiveness of small and large type 1 neurons to μ-opioid receptor agonists could be caused by density of μ-receptors, types of calcium channels involved (Moises et al. 1994), or differential efficacy of coupling between receptor and channel. Calcium channel types can be identified on the basis of their differential sensitivity to drugs and toxins from animal venoms. Application of a high concentration of the L-type calcium channel blocker nimodipine (3 μm) inhibited whole-cell HVA ICa in small type 1 sensory neurons by 7 ± 2 % (n= 11; Fig. 7A, G and H), large type 1 sensory neurons by 11 ± 1 % (n= 16; Fig. 7C and G), but did not inhibit ICa of type 2 sensory neurons (mean = 0.7 ± 0.8 %; n= 17, Fig. 7E and G). The block caused by nimodipine was reversible upon washing the cells (see Fig. 7C). The inhibition of whole-cell ICa by nimodipine was not affected by prior application of ω-GVIA (1 μm) or ω-Aga IVA (500 nm). Calcium channel blockers were applied in a counterbalanced order to reduce potential bias or artefacts caused by current run-down.

Figure 7.

HVA ICa are composed of different proportions of N-type, L-type, P/Q-type and R-type in the three subpopulations of mouse sensory neurons

A, C and E, time plots of the inhibition of HVA ICa by ω-GVIA, nimodipine and ω-Aga IVA in a small type 1 neuron, and large type 1 and type 2 neurons, respectively. B, D and F, example current-voltage relationships of a small type 1, and large type 1 and type 2 neurons, respectively. Triangles represent current resistant to ω-GVIA, nimodipine and ω-Aga IVA. Squares represent current in the absence of inhibitors. G, bars represent the percentage contribution of N-type (filled bars), P/Q-type (open bars), L-type (shaded bars) and R-type ICa (hatched bars) to the peak ICa for small type 1, and large type 1 and type 2 neurons. Each proportion represents data from at least 11 cells. H, example traces from a small type 1, N/OFQ-sensitive sensory neuron showing HVA ICa following application of ω-GVIA, nimodipine and ω-Aga IVA.

The N-type calcium channel blocker, ω-GVIA (1 μm), inhibited 52 ± 4 % (range = 22–81 %, n= 16; Fig. 7A, G and H) of whole-cell ICa in small type 1 neurons. Application of 1 μmω-GVIA inhibited whole-cell ICa by 43 ± 2 % (n= 15; Fig. 7C and G) in large type 1 and 34 ± 4 % (n= 17; Fig. 7E and G) in type 2 neurons. Out of the 16 type 2 neurons tested, two cells appeared to have no ω-GVIA-sensitive ICa.

Application of the P/Q-type calcium channel blocker, ω-Aga IVA (500 nm), produced a similar inhibition of ICa in all three cell types. The inhibition was 40 ± 4 % in small type 1 neurons (n= 16; Fig. 7A, G and H), 38 ± 4 % in large type 1 neurons (n= 15; Fig. 7C and G), and 44 ± 4 % in type 2 neurons (n= 12; Fig. 7E and G). Inhibition by ω-Aga IVA took several minutes to equilibrate and was not reversed by washing.

Most of the whole-cell current was blocked by simultaneous application of 3 μm nimodipine, 1 μmω-GVIA and 500 nmω-Aga IVA. However, some resistant current remained in each cell type, most notably in type 2 sensory neurons. The resistant component of small type 1 and large type 1 sensory neurons was 5 ± 1 % (n= 14; Fig. 7A, C, G and H) and 14 ± 1 % (n= 15, Fig. 7C, D and G), respectively. Type 2 sensory neurons had a resistant component of 28 ± 4 % (n= 13; Fig. 7E, F and G). The LVA ICa did not account for the 28 % resistant component because the latter was completely blocked by the non-selective HVA ICa blocker, Cd2+ (30 μm; Fig. 7E).

Modulation of subtypes of ICa by N/OFQ and DAMGO in small type 1 neurons

The inhibition of the various types of HVA ICa in small type 1 sensory neurons was examined by applying N/OFQ and DAMGO in the presence of combinations of calcium channel blockers in order to determine whether the smaller inhibition of HVA ICa produced by N/OFQ versus DAMGO was caused by selective modulation of different calcium channel subtypes.

Both DAMGO and N/OFQ selectively modulated N-type and P/Q-type currents. Neither DAMGO (1 μm; n= 6) nor N/OFQ (300 nm; n= 6) inhibited L/R-type ICa isolated in the presence of ω-GVIA and ω-Aga IVA toxins (data not shown). The inhibition of N/R-type ICa isolated in the presence of 3 μm nimodipine and 500 nmω-Aga IVA by DAMGO (1 μm) was significantly greater than the inhibition produced by N/OFQ (300 nm, P < 0.05, Fig. 8). DAMGO (1 μm) also inhibited P/Q/R-type ICa isolated with 3 μm nimodipine and 1 μmω-GVIA to a significantly greater extent than N/OFQ (300 nm; P < 0.005; Fig. 8). R-type currents, recorded in the presence of nimodipine, ω-GVIA and ω-Aga IVA were negligible in small type 1 neurons, and DAMGO (1 μm) or N/OFQ (300 nm) did not appreciably modulate isolated R current in large type 1 neurons (n= 4 each, data not shown). Therefore, it is reasonable to assume that inhibitions of N/R and P/Q/R currents by DAMGO and N/OFQ were caused by inhibition of N- or P/Q-type ICa, respectively.

Figure 8.

μ-Opioid- and ORL1-receptor agonists produce different maximal inhibitions of N/R and P/Q/R components of ICa

A, DAMGO (1 μm) produced a similar maximal inhibition of both N/R and P/Q/R ICa components. B, N/OFQ (300 nm) produced greater inhibition of N/R than P/Q/R ICa components (P < 0.01, unpaired Student's t test).


N/OFQ inhibited HVA ICa in a subpopulation of presumed nociceptive neurons by a voltage-sensitive, G-protein-dependent mechanism. The N/OFQ-sensitive neurons were small (calculated to be < 20 μm in diameter), μ-opioid and VR1 receptor agonist sensitive, and were labelled by IB4. Inhibition of HVA ICa by N/OFQ was potent and was blocked by the non-peptide ORL1 receptor antagonist, J 113397 (Ozaki et al. 2000), abolished by pre-treatment with PTX, associated with significant slowing of calcium channel activation and reduction of inhibition after a large depolarizing step in the presence of N/OFQ. The latter are characteristic features of Gi/o-protein βγ-subunit-mediated inhibition (Herlitze et al. 1996; Ikeda, 1996). N/OFQ has been reported to inhibit a variety of HVA ICa in a G-protein-dependent manner in several types of central (Knoflach et al. 1996; Connor & Christie, 1998) and peripheral neurons (Abdulla & Smith, 1997; Zhang et al. 1998; Larsson et al. 2000). As reported by Abdulla & Smith (1998) in rat DRG neurons, no inhibition of L-type calcium channels by N/OFQ was observed in the present study. Furthermore, no inhibition of LVA ICa by N/OFQ was observed in type 2 neurons, regardless of whether HVA ICa was inhibited in the same cells. G-protein-independent inhibition of LVA ICa by N/OFQ has been reported in a population of rat acutely isolated DRG neurons (Abdulla & Smith, 1997). The mechanism(s) by which ORL1 receptors inhibit LVA ICa in DRG neurons have not been defined, so we cannot say why this interaction was not observed in the present study, given that recording conditions were similar. It is possible that mouse trigeminal neurons express different types of LVA ICa, or co-express LVA ICa and insufficient amounts of ORL1 receptors or other signalling elements necessary for the interaction. Notably, the neurons in which N/OFQ inhibited the T-type currents were selected on the basis of cell size (70–90 pF) and the large amplitude of their LVA currents (Abdulla & Smith, 1997).

The contribution of HVA ICa types, as distinguished by selective channel blockers, was similar across the three trigeminal neuron classifications. However, type 2 sensory neurons had more R-type calcium current than type 1 sensory neurons. Interestingly, L-type calcium channels could not be detected in type 2 neurons. The presence of prominent LVA ICa in a well-defined population of trigeminal neurons was also striking. The LVA currents were more sensitive to Ni2+ than Cd2+, which suggests that they comprise, at least in part, α1H (CaV3.2)-containing channels (Lee et al. 1999). A prominent LVA ICa is associated with a characteristic burst-firing pattern in sensory neurons (White et al. 1989). Interestingly, a previous study in mouse DRG neurons also reported a population of neurons expressing prominent LVA ICa which were insensitive to capsaicin (Pearce & Duchen 1994); it will be of interest to see if these distinctive neurons share other common properties.

Supramaximal concentrations of DAMGO inhibited N- or P/Q-type channels equi-effectively in mouse trigeminal neurons. However, N/OFQ preferentially inhibited N-type calcium channels versus P/Q-type channels. Since receptor-mediated ICa inhibition is mediated primarily by G-protein βγ subunit association with calcium channel pore subunit (Herlitze et al. 1996; Ikeda, 1996), this differential inhibition of P/Q channels observed between DAMGO and N/OFQ may be indicative of preferential coupling of different G-protein β subunit isoforms to either receptor (Arnot et al. 2000). Alternatively, ORL1 and μ-opioid receptors may be differentially localized in the vicinity of different calcium channel types (e.g. Delmas et al. 2000). There is also kinetic evidence that various calcium channels have different affinities for βγ subunits (Arnot et al. 2000; Delmas et al. 2000), so it is possible that in small type 1 neurons μ-opioid receptor activation releases enough βγ subunits to saturate both N-type and P/Q-channels, while ORL1 receptor activation does not (perhaps because of lower receptor expression level).

Nearly all of the small type 1, N/OFQ-sensitive cells recorded in this study also responded to capsaicin, suggesting that they represent a subpopulation of VR1 expressing nociceptive neurons. In mouse, unlike rat (Guo et al. 1999), VR1 is usually co-localized with substance P and calcitonin gene related peptide in primary afferent fibres (Guo et al. 2000), suggesting that small and possibly some large type 1 neurons are peptide-containing nociceptors. IB4, which binds to most small type 1 neurons, is also thought to label a subpopulation of nociceptors that express P2X3 receptors (Guo et al. 1999) in rats, but appear not to in mice (Guo et al. 2000). Therefore, it is not yet clear whether small type 1 neurons are equivalent to the IB4-positive sensory neurons previously described in the mouse.

In a spinal cord slice preparation where the dorsal rootlets still attached to the dorsal horn were stimulated, N/OFQ inhibited only long-latency glutamatergic excitatory postsynaptic current amplitude in substantia gelatinosa neurons, a characteristic of C-fibre neurons (Luo et al. 2000). These data are consistent with N/OFQ inhibition of HVA ICa in small nociceptor-like trigeminal neurons.

Nearly all of the small type 1, N/OFQ-sensitive neurons recorded in this study also responded to μ-opioid agonists. In contrast, fewer large than small type 1 neurons responded to μ-opioids and the efficacy of opioid actions was lower in the former. These results are broadly consistent with the findings of Taddese et al. (1995) in rat nociceptive trigeminal ganglion neurons which innervate tooth pulp, suggesting that the small type 1 cells described here are indeed nociceptors. In the former study, DAMGO inhibited HVA ICa in 79 % of small (diameter < 30 μm) nociceptive neurons, but only inhibited HVA ICa in 11 % of large (diameter > 40 μm) nociceptors (Taddese et al. 1995).

The highly efficacious μ-agonist DAMGO was more potent and produced a greater maximum response in small compared to large type 1 neurons. Morphine, which has lower intrinsic efficacy at μ-receptors than DAMGO, had similar potency in both cell types, but produced a larger maximum response in small type 1 cells. These results suggest that the efficacy of coupling between μ-receptors and calcium channels is greater in small than in large type 1 neurons, assuming that there is little or no receptor reserve for morphine, but some for DAMGO in small type 1 neurons (e.g. Christie et al. 1987). More efficient coupling could result from higher expression of μ-receptors in small versus large trigeminal neurons, which has been demonstrated for μ-receptor mRNA expression in single nociceptive trigeminal neurons from rat (Silbert et al. 1998). The apparent differences in efficacy could also partly reflect differences in calcium channel subtypes since the toxin-typing strategy used in these experiments would probably not distinguish between different potential splice variants of α1A and α1B channels that appear to have different sensitivities to G protein βγ-subunit-mediated inhibition (Bourinet et al. 1999; Delmas et al. 2000).

HVA ICa inhibition by opioids in type 1 sensory neurons was mediated exclusively by μ-opioid receptors. Both μ- and κ-opioid agonists have been reported to inhibit HVA ICa in mouse sensory neurons (MacDonald & Werz 1986; Jiang et al. 1998), but extensive studies with more selective opioid agonists have been restricted to rat neurons thus far. In rat sensory neurons, μ-, δ- and κ-receptor agonists have been shown to inhibit ICa (Schroeder et al. 1991; Moises et al. 1994; Acosta & Lopez, 1999), although only in trigeminal ganglion neurons has a strong correlation between μ-opioid inhibition of ICa and nociceptive modality been established (Schroeder & McClesky, 1993; Taddese et al. 1995; Su et al. 1998).

N/OFQ had little or no action on type 2 neurons. Although they were relatively small, almost all type 2 neurons were insensitive to capsaicin and none responded to μ-opioids. We have not established the sensory modality of this population of cells. It is possible that they are nociceptors which respond to stimuli not uniquely transduced by polymodal nociceptors, such as cold (Gallar et al. 1993) and changes in pH (Chen et al. 1997), or it is possible that they are not nociceptive neurons. Few type 2 neurons responded to N/OFQ and those that did responded weakly. Therefore, the present results suggest that ORL1 receptor agonists should have anti-nociceptive actions in the sensory trigeminal system and are consistent with inhibition of C-fibre evoked responses and spinal anti-nociceptive actions of N/OFQ (see ‘Introduction’). However, it remains unclear how the pro-nociceptive actions of N/OFQ in mice (Inoue et al. 1998) and rats (Carpenter et al. 2000) could be mediated by actions of this inhibitory Gi/o coupled receptor in the peripheral terminals of small, nociceptive neurons.

In conclusion, the present studies have established that N/OFQ selectively inhibits HVA ICa in a population of small, presumably nociceptive, but not large VR1- and μ-opioid receptor-expressing nociceptive trigeminal ganglion neurons that do not express LVA ICa (type 1 neurons). In contrast, small neurons that express LVA ICa usually do not respond to VR1 or μ-opioid receptor agonists, and are relatively insensitive to N/OFQ (type 2 neurons). It is not yet known whether type 2 neurons represent a different subpopulation of nociceptors. Therefore, N/OFQ is likely to have anti-nociceptive actions at the central terminals of trigeminal neurons. Co-expression of ORL1, μ-opioid and VR1 receptors, together with absence of LVA calcium channels, may be also used to define a subpopulation of small nociceptive neurons in the trigeminal system.


This research was supported by the National Health and Medical Research Council of Australia (grants 153911 and 107489), and The Medical Foundation of the University of Sydney. We are grateful to Banyu Pharmaceuticals Inc. for supplying us with J-113397.